Meet the Adgent: Your AI Analyst, Built Right Into AdCrunch
AdCrunch’s console now has its own AI chat agent. Ask about your campaigns in plain language and get answers grounded in your organization’s live data — no external AI client, no copy-pasting context, no leaving the product.
- Product update
- MCP
You already connected AdCrunch to Claude, Cursor, or ChatGPT over MCP — that’s still the fastest way to hand an AI agent your live ad data, and that isn’t changing.
But every time you want to ask something, you’re opening a different app. Re-explaining which organization you mean. Starting from a blank chat that knows nothing about the conversation you had yesterday.
What if the analyst just lived where your campaigns already live?
Meet the Adgent
The Adgent is AdCrunch’s own AI chat agent, built directly into the console. Sign in, open it, and ask:
“How did my Meta campaigns do last week?”
No API key to paste in. No MCP server to configure. No tab-switching. The Adgent already knows who you are and which organization you’re looking at — it answers using your live Insights, grounded in real numbers, not a guess.
It’s Not a Bolted-On Chatbot
Here’s the part that matters: the Adgent doesn’t have its own, separate, thinner view of your data. It’s built as an MCP client of AdCrunch’s own server — the exact same tools an external agent uses when you connect Claude or Cursor to query your campaigns, ad sets, and performance.
That has real consequences for what it can do:
| Without the Adgent | With the Adgent |
|---|---|
| Open a separate AI app to ask a question | Ask right inside the console |
| Re-explain which org or account you mean | Already scoped to your active organization |
| Rebuild context every session | Your conversation history is already there |
| Hope you remember your team’s saved prompts | Runs your saved Skills automatically |
| A different tool than what your team uses externally | Same tools, same data, same answers — just a different door |
If your team has a weekly-meta-audit Skill saved, the Adgent runs it the same way your external agent would — just ask for it by name. And because it’s calling the same live tools, it handles the same edge cases: if your advertisers bill in more than one currency, it converts to a single display currency instead of silently mixing totals.
Built for How You Actually Work
Multiple conversations, not one long scroll. Start a new thread for a new topic, and switch back to an old one from the panel on the left. Each conversation keeps its own history and picks up a short, auto-generated title after your first message — so your list stays scannable instead of a wall of “New conversation.”
It streams. Answers appear token by token, like any chat interface you’re used to — not a spinner followed by a wall of text.
It picks up where you left off. Close the tab, come back tomorrow, switch devices — your conversations are right there.
It’s private, per person, per organization. Nobody else in your org can read your chats. If you belong to more than one organization, each one gets its own separate history — switching your active org never mixes one client’s campaigns into another’s conversation.
It won’t touch your campaigns. Today, the Adgent is read-only. It can report on performance, discover your advertisers and entities, and run your Skills — but it cannot pause a campaign, edit a budget, or change anything. Ask it anything you want; it can’t break anything.
Try It
If you’re already signed in to AdCrunch, the Adgent is one click away in the console nav — nothing to connect, nothing to configure. Ask it your weekly numbers, a campaign that’s been bugging you, or anything else you’d normally build a report to answer.
New to AdCrunch?